The Master Neural Transcription Factor BRN2 Is an Androgen Receptor–Suppressed Driver of Neuroendocrine Differentiation in Prostate Cancer

Published OnlineFirst October 26, 2016; DOI: 10.1158/2159-8290.CD-15-1263

RESEARCH ARTICLE

Jennifer L. Bishop1, Daksh Thaper1,2, Sepideh Vahid1,2, Alastair Davies1, Kirsi Ketola1, Hidetoshi Kuruma1, Randy Jama1, Ka Mun Nip1,2, Arkhjamil Angeles1, Fraser Johnson1, Alexander W. Wyatt1,2, Ladan Fazli1,2, Martin E. Gleave1,2, Dong Lin1, Mark A. Rubin3, Colin C. Collins1,2, Yuzhuo Wang1,2, Himisha Beltran3, and Amina Zoubeidi1,2

ABSTRACT Mechanisms controlling the emergence of lethal neuroendocrine prostate cancer (NEPC), especially those that are consequences of treatment-induced suppression of the androgen receptor (AR), remain elusive. Using a unique model of AR pathway inhibitor–resistant prostate cancer, we identified AR-dependent control of the neural transcription factor BRN2 (encoded by POU3F2) as a major driver of NEPC and aggressive tumor growth, both in vitro and in vivo. Mecha- nistic studies showed that AR directly suppresses BRN2 transcription, which is required for NEPC, and BRN2-dependent regulation of the NEPC marker SOX2. Underscoring its inverse correlation with classic AR activity in clinical samples, BRN2 expression was highest in NEPC tumors and was significantly increased in castration-resistant prostate cancer compared with adenocarcinoma, especially in patients with low serum PSA. These data reveal a novel mechanism of AR-dependent control of NEPC and suggest that targeting BRN2 is a strategy to treat or prevent neuroendocrine differentiation in prostate tumors.

SIGNIFICANCE: Understanding the contribution of the AR to the emergence of highly lethal, drug- resistant NEPC is critical for better implementation of current standard-of-care therapies and novel drug design. Our first-in-field data underscore the consequences of potent AR inhibition in prostate tumors, revealing a novel mechanism of AR-dependent control of neuroendocrine differentiation, and uncover BRN2 as a potential therapeutic target to prevent emergence of NEPC. Cancer Discovery; 7(1); 1–18. ©2016 AACR.

1Vancouver Prostate Centre, Vancouver, British Columbia, Canada. Corresponding Author: Amina Zoubeidi, Vancouver Prostate Centre, 2660 2Department of Urologic Sciences, Faculty of Medicine, University of Oak Street, Vancouver, BC V6H3Z6, Canada. Phone: 604-809-4947; Fax: British Columbia, Vancouver, British Columbia, Canada. 3Department of 604-875-5654; E-mail: [email protected] Medicine, Division of Hematology and Medical Oncology, Weill Cornell doi: 10.1158/2159-8290.CD-15-1263 Medical College, New York, New York. ©2016 American Association for Cancer Research. Note: Supplementary data for this article are available at Cancer Discovery Online (http://cancerdiscovery.aacrjournals.org/).

OF1 | CANCER DISCOVERY JANUARY 2017 www.aacrjournals.org

INTRODUCTION for neuroendocrine (NE) markers that characterize NEPC, it is often distinguished from prostatic adenocarcinoma by Progression from primary prostate cancer to advanced reduced AR expression or activity (5). Clinical presentation metastatic disease is heavily dependent on the androgen of NEPC reflects this shift away from reliance on the AR, as receptor (AR), which fuels tumor survival. In men whose patients typically present with low circulating levels of PSA treatments for localized prostate tumors have failed, or despite high metastatic burden in soft tissues, and are refrac- in those who present with metastatic disease, androgen tory to APIs (3). Importantly, it has been reported that under deprivation therapies (ADT) are used to deplete circulating the strong selective pressure of potent APIs like ENZ, these androgens to abrogate AR signaling and prevent disease “non–AR-driven” prostate cancers, which include NEPC, progression. Eventually, however, prostate cancer recurs may constitute up to 25% of advanced, drug-resistant CRPC after first-line ADT as castration-resistant prostate cancer cases (6). Not surprisingly, therefore, the incidence of NEPC (CRPC). Despite low levels of serum androgens in men with has significantly increased in recent years (7), coinciding with CRPC, reactivation of the AR occurs; thus, it remains central the widespread clinical use of APIs. to tumor cell survival, proliferation, and metastatic spread. A number of molecular mechanisms likely facilitate the Targeting the AR is a cornerstone therapeutic intervention progression of CRPC to NEPC. These include loss of tumor in patients with CRPC, and AR pathway inhibitors (API) suppressors, such as RB1 (8, 9) and p53 (10), amplification that further prevent AR activation, such as enzalutamide of MYCN (11), mitotic deregulation through AURKA (11) (ENZ), have become mainstays in the prostate cancer treat- and PEG10 (12), epigenetic controls such as REST (13–15) ment landscape (1). Despite its being a potent API, the treat- and EZH2 (11, 16), and splicing factors like SSRM4 (14, 17). ment benefits of ENZ are short-lived in patients with CRPC Importantly, the AR plays a crucial, albeit still mechanisti- and resistance rapidly occurs (2). cally unclear, role in NEPC. Reports over many years have ENZ-resistant (ENZR) CRPC represents a significant clini- highlighted how ADT (18, 19) or loss of AR promotes the NE cal challenge not only due to the lack of third-line treatment differentiation of prostate cancer cells (reviewed in ref. 20); as options to prevent AR-driven tumor progression but also such, many genes associated with an NE phenotype, includ- because it can be a precursor to rapidly progressing and ing ARG2 (21), HASH1 (22), and REST (14, 15) are controlled lethal neuroendocrine prostate cancer (NEPC). Although by the AR. Although this evidence underscores an inverse NEPC can rarely arise de novo, it is increasingly defined as correlation between AR expression and/or activity and mol- a variant of highly API-resistant CRPC (3, 4). Aside from ecular events leading to NEPC, the mechanisms by which the the unique small-cell morphology and positive staining AR directly influences the induction of an NEPC phenotype

JANUARY 2017 CANCER DISCOVERY | OF2

RESEARCH ARTICLE Bishop et al. from CRPC under the selective pressure of APIs such as ENZ 42DENZR, 42FENZR, 49CENZR, 49FENZR, etc.; Supplementary remain elusive. Fig. S1A–S1C) and were screened for protein expression of Answering such questions requires a model of API-resist- AR and PSA. Reflectingin vivo data, the established cell lines ant CRPC that recapitulates the transdifferentiation of ade- displayed heterogeneous expression of PSA, yet all retained nocarcinoma to NEPC that occurs in patients. Herein, we expression of the AR (Fig. 1G). Importantly, 42DENZR present an in vivo–derived model of ENZR, which, different and 42FENZR cell lines derived from PSA- tumors (Fig. 1E from others (23–25), underscores the emergence of tumors and Supplementary Fig. S1C and S1D) remained PSA- in vitro with heterogeneous mechanisms of resistance to ENZ over (Fig. 1G), whereas 49CENZR and 49FENZR cells derived from multiple transplanted generations. These include the natu- PSA+ tumors (Fig. 1F; Supplementary Fig. S1B and S1D) ral acquisition of known AR mutations found in ENZR retained PSA expression (Fig. 1G). Accordingly, sequenc- patients (25–28) and the transdifferentiation of NEPC-like ing the ligand binding domain (LBD) of the AR revealed tumors through an AR+ state, without the manipulation of the presence of the F878L AR activating mutation in PSA+ oncogenes typically used to establish NEPC in murine pros- 49CENZR and 49FENZR cells but not in PSA− 42DENZR and tate cancer models (29–31). Using this model and data from 42FENZR cells (Fig. 1H). Emergence of this mutation in only patients with prostate cancer, we show that a master regula- PSA+ ENZR cells supports previous results showing this tor of neuronal differentiation, the POU-domain transcrip- alteration mediates resistance to ENZ (25–28) and suggests tion factor BRN2 (encoded by POU3F2; ref. 32), is directly that it may be one mechanism by which the AR is reactivated transcriptionally repressed by the AR, is required for the specifically in this subset of ENZR cells. expression of terminal NE markers and aggressive growth Mounting evidence suggests that “non–AR-driven” phe- of ENZR CRPC, and is highly expressed in human NEPC notypes of prostate cancer can emerge under the selective and metastatic CRPC with low circulating PSA. Beyond pressure of potent APIs like ENZ. Importantly, “non–AR- suppressing BRN2 expression and activity, we also show driven” tumors are not necessarily negative for AR expression; that the AR inhibits BRN2 regulation of SOX2, another in fact, many retain AR but show reduced AR activity (33). transcription factor associated with NEPC. These results Accordingly, our model of ENZR showed that approximately suggest that relief of AR-mediated suppression of BRN2 is a 25% of serially transplanted tumors, which emerged under consequence of ENZ treatment in CRPC that may facilitate constant presence of ENZ, may be non–AR-driven (Fig. 1D), the progression of NEPC, especially in men with “non–AR- a distribution that is echoed in clinical reports (4, 6). Indeed, driven” disease. RNA sequencing (Fig. 1I) and microarray analysis (Sup- plementary Fig. S1E) of PSA- ENZR cells indicated that RESULTS these cells exhibited divergence in global RNA expression on principal coordinate analysis (PoCA) from either CRPC or Emergence of AR-Driven and Non–AR-Driven PSA+ (“AR-driven”) cells and displayed reduction not only R Tumors in ENZ CRPC in PSA (encoded by KLK3) but in many classic AR-regulated To model ENZR disease, we developed cell lines from genes compared with 16DCRPC cells (Fig. 1J; Supplemen- LNCaP-CRPC and ENZR LNCaP-CRPC xenograft tumors. tary Fig. S1F and Supplementary Table S1). Underscor- LNCaP cells were used to establish subcutaneous tumors ing the AR dependency of AR-driven versus non–AR-driven in intact male athymic nude mice and, upon tumor growth ENZR phentoypes, targeting AR using siRNA significantly and rising PSA, mice were castrated. Once tumors recurred reduced cell proliferation in 49FENZR cells, whereas no effect (CRPC), mice were treated with vehicle or 10 mg/kg ENZ was observed in 42DENZR and 42FENZR cells (Supplementary daily and monitored for tumor growth (Supplementary Fig. S1F). These data suggest that emergence of ENZ resis- Fig. S1A and Supplementary Methods). Although ENZ tance does not require AR reactivation and that our ENZR treatment did slow tumor growth compared with vehicle model may have utility in studying mechanisms of both clas- control, it did not prevent tumor recurrence (Fig. 1A) and sic AR-driven and non-driven disease phenotypes, especially the majority (9 of 10, 90%) of ENZ-treated CRPC tumors transdifferentiation of AR+ ENZR CRPC to NEPC. increased in tumor volume with concomitant rise in PSA (Fig. 1B). Importantly, however, PSA was not required for The Neural Transcription Factor BRN2 Is Highly R tumor recurrence, as observed in one mouse (Fig. 1B and C). Expressed in NE-like ENZ and in Human NEPC PSA+ CRPC tumors that grew in the presence of ENZ were Our data analysis revealed that 42DENZR and 42FENZR serially transplanted into castrated male mice treated with cells exhibited reduced expression of classic AR target genes 10 mg/kg ENZ to establish ENZR tumors (Supplementary (Fig. 1J; Supplementary Fig. S1G). Additionally, 42DENZR Fig. S1A). Similar to the primary CRPC parental xenografts and 42FENZR cells showed increased expression of canonical that recurred in the presence of ENZ, the majority (26 of transcription factors and markers associated with neuronal 35, 74.3%) of transplanted tumors showed increasing vol- development and NEPC (5), such as neuron-specific eno- ume associated with rising PSA (Supplementary Fig. S1B). lase (NSE), synaptophysin (SYP), chromagranin A (CGA, However, over the course of serial transplantation, primary encoded by CHGA), and neural cell adhesion marker 1 PSA+ ENZR xenografts also gave rise to 9 serially trans- (NCAM1; Fig. 2A; Supplementary Fig. S2A and Supplemen- planted tumors out of 35 (25.7%) that grew without rise in tary Table S1). The increased expression of each of these PSA (Fig. 1D–F; Supplementary Fig. S1C). Cell lines were terminal NE markers was validated at the mRNA (Fig. 2B derived from vehicle-treated CRPC (referred to as 16DCRPC) and C) and protein levels (Fig. 2D) in 42DENZR and 42FENZR and multiple transplanted ENZR tumors (referred to as cell lines and in tumors compared with CRPC controls.

OF3 | CANCER DISCOVERY JANUARY 2017 www.aacrjournals.org

BRN2 Is an AR-Suppressed Driver of NEPC RESEARCH ARTICLE

AB C Tumor volume PSA Primary 600 R ) 2,500 ENZ xenografts 3 CRPC CRPC R R 500 ENZ +PSA 2,000 ENZ R ENZR−PSA ENZ +PSA 400 ENZR−PSA 1,500 300 1 1,000 +ENZ/V n = 10 200

500 (ng/mL) PSA 100

Tumor volume (mm Tumor 9 0 0 −4 −2 0 2468 −4 −2 0 2468 Post-Cx Tx Post-Cx Tx Time (weeks) Time (weeks) N = 10 DE F R R Transplanted ENZ transplant #42 ENZ transplant #49 R ENZ + PSA xenografts 2,500 500 ) 2,500 500 ) 3 Volume 3 Volume ENZR+PSA PSA (ng/mL) ENZR−PSA 2,000 PSA 400 2,000 PSA 400 PSA (ng/mL) 9 1,500 300 1,500 300 1,000 200 1,000 200 26 500 100 500 100 Tumor volume (mm Tumor 0 0 volume (mm Tumor 0 0 0246810 12 14 16 02468 N = 35 Time (weeks) Time (weeks)

GH AR mutations 100 22RV1 CRPC ENZR ENZR ENZR ENZR ENZR ENZR ENZR ENZR LAPC4 75 16D 42D 42F 48E 48F 49B 49C 49D 49F F877L T878A LNCaP IARELHQFTFDLLIKSHMVSVDFPEMMAEIISVQVPKILSGKVK CRPC AR 16D NTD DBD LBD 50 49FENZR ENZR 12345678 49C PSA 25 ENZR AR mutated (%) 49D VINC 0 T878A F877L H874Y IJCell line principal coordinates analysis AR targets 0.010 ENZR 49F 1.3x 0.1x 0.005 ENZR 49C CRPC CRPC 0.000 16D 16D

ENZR F2 (12.3%) CRPC −0.005 16A 42D P ENZR PA PSA 42D PrLZ PSCA PSMA PSGR AlbZIP hKLK4 hKLK2 FKBP5 STEAP PSDR1 NKX3.1 Prostein

−0.010 PMEPA1

0.03 0.02 0.01 0.00 0.01 0.02 0.03

F1 (81.3%)

Figure 1. Emergence of AR-driven and non–AR-driven phenotypes in an in vivo model of ENZR CRPC. A and B, 1 × 106 LNCaP cells were used to establish primary and CRPC subcutaneous xenografts in male athymic nude mice. Graphs show tumor volume and corresponding serum PSA levels of tumors from 4 weeks after castration (Cx). Time 0 represents time at which serum PSA levels reached precastration levels and start of treatment (Tx) with vehicle (V, n = 10) or 10 mg/kg ENZ (ENZ, n = 10). A, Volumes of 10 vehicle-treated CRPC (black line) and 10 ENZ-treated (ENZR; purple line) tumors. B, Black line shows average serum PSA of vehicle-treated mice, blue line represents average serum PSA of 9 out of 10 tumors that recurred in the presence of ENZ (ENZR+PSA) and red line represents average serum PSA of remaining 1 tumor that recurred in the presence of ENZ (ENZR - PSA; n = 10 total, represented in purple line from A). C and D, Fraction of total (C) primary, or (D) transplanted, ENZR xenografts that recurred with (blue) or without (red) parallel rise in serum PSA (ENZR+/−PSA). E and F, Tumor volume and serum PSA levels for individual transplanted ENZR xenografts (#42 and #49) used to derive ENZR cell line clones 42D and 42F or 49C and 49F, respectively. G, Protein expression of AR, PSA, and vinculin (VINC) in CRPC and ENZR cell lines. H, Percent AR mutations in 22RV1, LAPC4, LNCaP, CRPC, and ENZR cell lines. I, Differences in global RNA expression were determined using multidimensional PoCA of RNA-sequencing data from CRPC and ENZR cell lines. J, Heat map showing fold increase in RNA-sequencing reads per million of AR target genes in 42DENZR cells compared with 16DCRPC (= 1). See also Supplementary Fig. S1 and Supplementary Table S1.

JANUARY 2017 CANCER DISCOVERY | OF4

RESEARCH ARTICLE Bishop et al.

Significantly increased surface expression of NCAM1 on Increased BRN2 expression in NEPC was also observed by 42DENZR and 42FENZR cells compared with 16DCRPC cells IHC staining (Fig. 3G and H). Importantly, IHC analysis of was confirmed by flow cytometry (Fig. 2E). Importantly, not only human NEPC but also CRPC and adenocarcinoma we found that BRN2 was strongly upregulated in our RNA supported the inverse correlation between BRN2 and AR sequencing data in 42DENZR versus 16DCRPC cells (Fig. 2A) (Pearson R = −0.144, P = 0.0043) and positive correlations as well as in microarray data comparing 42FENZR versus with CGA (Pearson R = 0.2686, P < 0.0001) and SYP (Pearson 16DCRPC cells (Supplementary Fig. S2A). Across multiple R = 0.2709, P < 0.0001). Furthermore, in patients, BRN2 was prostate cancer cell lines, BRN2 mRNA expression (Fig. 2F) more highly expressed in metastatic CRPC than in localized was highest in the NEPC tumor–derived NCIH660 cell line adenocarcinoma (ref. 37; Supplementary Fig. S3A) and in and second highest in 42DENZR and 42FENZR cells. These metastatic than primary prostate cancer (ref. 38; Supplemen- data were reflected in serially transplanted xenografts that tary Fig. S3B). Finally, expression of BRN2 positively cor- gave rise to 42DENZR and 42FENZR cells but not primary related with the NE-associated genes CGA, CGB, SYP, and CRPC tumors (Supplementary Fig. S2B). These ENZR cells MYCN (ref. 38; Supplementary Fig. S3C). These data show showed the highest BRN2 activity (Fig. 2G), and increased for the first time that BRN2 expression is strongly associ- BRN2 expression was validated at the protein level by ated with severity of disease in prostate cancer, especially an Western blot (Fig. 2H) and IHC (Supplementary Fig. S2C) NE phenotype, and that it is inversely correlated with AR in 42DENZR and 42FENZR cells and tumors compared with activity. 16DCRPC. In addition to NCIH660 and our ENZR cell lines, BRN2 expression was also increased in NE-like TRAMP+ BRN2 Is Inversely Correlated with AR Expression transgenic prostate tumors (34, 35), where it positively cor- and Activity related with the expression of NSE, SYP, and CGA (Fig. 2I Our observations in both NEPC patient tumors and our and J). These data suggest that non–AR-driven models of ENZR model led us to test the hypothesis that inhibition of ENZR display NE characteristics that may be supported by the classical AR pathway increases the expression of BRN2. the expression of BRN2. In accordance with our in vitro model where we observed To investigate the clinical relevance of BRN2 in human PSA− cell lines express BRN2, analysis of the CRPC samples NEPC, we assessed BRN2 expression in RNA sequencing in Grasso and colleagues (37) as well as the prostate adeno- data from prostatic hormone-naïve adenocarcinoma, CRPC, carcinoma data from The Cancer Genome Atlas showed an and NEPC patient tumors (Beltran cohorts; 2016, ref. 33, inverse correlation between BRN2 and serum PSA (Fig. 4A and 2011, ref. 11). In NEPC tumors, characterized by a high and Supplementary Fig. S3D). This trend was mirrored NEPC score and upregulation of canonical NE genes (Fig. 3A by immunohistochemistry analysis of a tissue microarray and B), BRN2 expression was significantly increased com- (TMA) of both CRPC specimens and treatment-naïve adeno- pared with CRPC or adenocarcinoma (Fig. 3C). Beyond this carcinoma, where we found a significant inverse correlation gene expression profile, NEPC is often associated with low between BRN2 staining intensity and circulating PSA levels AR activity; in these patient tumors, the AR activity score was in primary and CRPC patients. Moreover, BRN2 staining significantly lower than adenocarcinoma or CRPC (Fig. 3D intensity significantly increased in progression from pri- and E). Importantly, the AR score also decreased in CRPC mary prostate cancer to CRPC only in patients with low compared with adenocarcinoma, whereas BRN2 was sig- levels of circulating PSA (Fig. 4B). In vitro studies further nificantly increased (Fig. 3C and D), suggesting that BRN2 indicated that suppression of AR signaling regulates BRN2. expression may be associated not only with NEPC but also RNA sequencing (Supplementary Fig. S4A), Western blot with prostate cancer progression after ADT. Indeed, BRN2 analysis (Supplementary Fig. S4B), and IHC (Supplemen- expression and activity as well as NE marker expression tary Fig. S2C) showed that, compared with 16DCRPC, the were also increased in 16DCRPC compared with parental AR-driven, PSA+ ENZR cell lines did not upregulate BRN2 LNCaP cells (Fig. 2F and G; Supplementary Fig. S2D), data or markers of NE differentiation and showed significant that were in accordance with our observation in a human reduction in BRN2 activity (Supplementary Fig. S4C). CRPC patient-derived xenograft (PDX) that transdifferen- Serially transplanted tumors that gave rise to 49F cells tiated to NEPC in castrated mice (36). In this model, we (Supplementary Fig. S1B) were also negative for BRN2 and found that BRN2 expression was highly upregulated in CGA (Supplementary Fig. S4D). Notably, in 16DCRPC cells, transdifferentiated NEPC versus adenocarcinoma (Fig. 3F). BRN2 protein expression was increased after 2 days of ENZ

Figure 2. Non–AR-driven ENZR cells display an NE differentiation signature and increased levels of the neural transcription factor BRN2. A, Heat map showing fold increase in reads per million of genes involved in NE differentiation in 42DENZR cells compared with 16DCRPC (= 1). B and C, Relative mRNA expression of (B) NSE, SYP, CGA, and (C) NCAM1 in 42DENZR and 42FENZR cells compared with 16DCRPC (= 1). D, Protein expression of CGA, NSE, SYP, and VINC in LNCaP, 16DCRPC, 42DENZR, and 42FENZR cells and tumors. E, Frequency of live (7-AAD−) NCAM1+ in LNCaP, 16DCRPC, 42DENZR, and 42FENZR cells. F, Relative mRNA expression of BRN2 in prostate cancer (PCa) cell lines compared with LNCaP (= 1). G, Relative activity of luciferase under the control of BRN2 48 hours after transfection in prostate cancer cells compared with LNCaP (= 1). Luciferase (Luc) activity of BRN2 is normalized to Renilla. H, Protein expression of BRN2 and VINC in LNCaP, 16DCRPC, 42DENZR, and 42FENZR cells (top) and LNCaPnaïve, 16DCRPC, and 42DENZR tumors (bottom). I, RPKM score of BRN2 in prostate tumors or tissue isolated from TRAMP+ transgenic mice compared with normal. J, Pearson score of NSE, SYP, and CGA compared with BRN2 in TRAMP+ prostates. See also Supplementary Fig. S2 and Supplementary Table S1. Statistical analyses were performed on pooled data from at least three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

OF5 | CANCER DISCOVERY JANUARY 2017 www.aacrjournals.org

BRN2 Is an AR-Suppressed Driver of NEPC RESEARCH ARTICLE

A B C NE differentiation NE markersNCAM1 15 CRPC 950 180× 1× 16D 800 *** ** ENZR 42D 650 CRPC * 16D 10 42FENZR 500 350 ENZR 42D * 200 *** 5 50 *** * 1.0 SYP NSE * GCG TFF1 BRN2 MYT1 CHGA SCGN CHGB

MYCN 0.5 AURKB AURKA NKX2.2 NCAM1 Relative expression Relative expression 0.0 0 CRPC ENZR ENZR NEUROG1 NSE SYP CGA 16D 42D 42F

DE CRPC ENZR ENZR

LNCaP 16D 42D 42F LNCaP 16DCRPC NSE

SYP

CGA *** 3.2% 14.1% VINC ±1.5 ±0.59

Cell lines 7-AAD 42DENZR 42FENZR CRPC ENZR naïve LN 16D 42D

NSE *** *** SYP 51.6% 43.1% ±7.46 ±5.32 CGA NCAM1 VINC Tumors FGH BRN2 expression BRN2–Luc activity in PCa cell lines in PCa cell lines

6 ** CRPC ENZR ENZR CRPC ENZR 2,300 **** naïve 5 LNCaP 16D 42D 42F LN 16D 42D 2,000 4 * ** BRN2 125 **** 75 3 *** VINC 25 2 15 *** Cell lines Tumors ** 1 5 Relative luc activity Relative expression 0 0

C42 PC3 PC3 VCaP C4–2 LAPC4 LNCaP CPRC DU145 LAPC4 LNCaP CRPCDU145 42FENZR42DENZR 42FENZR42DENZRNCIH660

IJ BRN2 expression in BRN2 vs. NE markers in TRAMP model 10,000 TRAMP+ prostates 1,000 1,000 Marker Pearson R P

100 NSE0.6920 0.0389* 100 SYP 0.6216 0.0739 10 CGA 0.8290 0.0057** RPKM (Log10) BRN2 RPKM (Log10) 10 1 Normal TRAMP+ 10 100 1,000 NE markers RPKM (Log10)

JANUARY 2017 CANCER DISCOVERY | OF6

RESEARCH ARTICLE Bishop et al.

AB C BRN2 expression Beltran, 2016 Beltran, 2011 human PCa NEPC score Adeno CRPC NEPC Adeno NEPC **** 1.5 **** AURKA **** **** AURKB 6 * * CHGA 1.0 CHGB 5 DNMT1 NSE 4 MYCN 0.5 MYT1 3 NCAM1 NEUROG1 2 0.0 NKX2-2 1 SCGN scaled to mean SYP Log2 (RPKM + 1) 0 Integrated NEPC score −0.5 TFF1 Adeno CRPC NEPC Adeno CRPC NEPC

DEBeltran, 2016 Beltran, 2011 F

Adeno CRPC NEPC Adeno NEPC BRN2 expression AR score ABCC4 ACSL3 in PDX ADAM7 AMACR 0.8 **** AR 1,000 **** CAMKK2 0.6 CENPN C1orf116 ELL2 0.4 FKBP5 FOLH1 100 0.2 GNMT KLK2 10 0.0 KLK3 NKX3-1 −0.2 PAP read counts PSCA AR signaling score 0.4 PMEPA1 1 − PSMA1 Normalized RNA-Seq Adeno CRPC NEPC STEAP1 Adeno NEPC SLC45A3 TMPRSS2

GH BRN2 AR CGA

BRN2 expression IHC 2.25 NEPC

2.00

1.75

1.50 CRPC IHC score 1.25

1.00 Adeno CRPC NEPC Adeno

Figure 3. BRN2 is highly expressed in human NEPC. A, NEPC score (33); B, heat map of neuroendocrine associated genes, C, BRN2 expression (line at mean), D, AR score (11) and E, heat map of AR-regulated genes in RNA-sequencing data, obtained from two cohorts (11, 33) of Adeno (n = 98), CRPC (n = 32), and NEPC (n = 21) patients. F, BRN2 expression in patient-derived prostatic adenocarcinoma xenografts (Adeno) and terminally transdifferentiated NEPC tumors (NEPC) based on RNA sequencing. G, IHC score for BRN2 protein expression in adenocarcinoma (Adeno, n = 93), CRPC (n = 30), and NEPC (n = 11). H, Representative IHC for AR, BRN2, and CGA in Adeno, CRPC, and NEPC tumors. See also Supplementary Fig. S3 and Supplementary Table S2. Statistical analyses were performed on pooled data from at least three independent experiments. *, P < 0.05; ****, P < 0.0001. PCa, prostate cancer. treatment, which was followed by increased expression of the LAPC4, and 49FENZR cells was sufficient to induce expression terminal NE markers CGA, NSE, and SYP over 7 days of ENZ of NSE, CGA, SYP, and NCAM1 (Supplementary Fig. S5A– treatment (Fig. 4C). Importantly, siRNA-mediated silenc- S5D) and enriched for an NEPC gene signature and neuronal ing of BRN2 over the course of ENZ treatment prevented associated pathways in BRN2-overexpressing 16DCRPC cells the ENZ-induced upregulation of NE markers in 16DCRPC (Supplementary Fig. S5E; Supplementary Tables S2 and S3). cells (Fig. 4C) and in LAPC4 cells (Supplementary Fig. S4E). Moreover, the effect of BRN2 overexpression on NE markers Similarly, deletion of BRN2 by CRISPR/Cas9 gene editing was enhanced with ENZ treatment of 16DCRPC cells (Fig. 4E). (Fig. 4D) or stable knockdown by shRNA (Supplementary Importantly, overexpression of BRN2 in not only 16DCRPC Fig. S4F) in 16DCRPC cells prevented ENZ-induced upregula- cells (Fig. 4F) but also AR-driven 49FENZR cells (Supplemen- tion of NEPC markers, further confirming that BRN2 is a tary Fig. S5F) reciprocally downregulated AR target gene prerequisite for terminal NE marker expression. Recipro- expression, which was associated with significantly reduced cally, transient overexpression of BRN2 in 16DCRPC, PC3, sensitivity to ENZ in both cell lines (Fig. 4G; Supplementary

OF7 | CANCER DISCOVERY JANUARY 2017 www.aacrjournals.org

BRN2 Is an AR-Suppressed Driver of NEPC RESEARCH ARTICLE

AB Primary CRPC BRN2 vs. serum PSA BRN2 vs. serum PSA

(Grasso et al. 2012) 2.5** * 1 ** 2.0

0 1.5 PSA 0–10 PSA 1.0 −1 IHC score Pearson = −0.375 BRN2 mRNA P = 0.0445 0.5 −2 0.0 0510 15 Primary CRPC

Serum PSA (ng/mL) PSA 0–10 ng/mL 40–50 PSA PSA 40–50 ng/mL

C ENZ-induced NE ENZ-induced NE 20 siCTR + ENZ siBRN2 + ENZ * 16DCRPC + ENZ siBRN2 + ENZ 02 47 02 47 15 BRN2 BRN2 NSE * * SYP NSE * 10 * CGA SYP NCAM *

CGA Relative expression 5 * * * * VINC * * * *

0 02d4d7d02d 4d 7d

D Pou3f2 ch 6q16.1 ENZ-induced NE Exon 1 25 WT ENZ (7 days) WT + ENZ *** BRN2 KO Cln #3 + ENZ WT BRN2 KO WT WT #3 #7 20 Cln #7 + ENZ BRN2 ** 15 ** Cln#3 NSE * SYP 10 WT CGA Relative expression 5 VINC Cln#7 0 NSECGA SYP NCAM

EF G BRN2 overexpression AR targets 16DCRPC response to ENZ 16DCRPC + ENZ 16DCRPC + ENZ + ENZ + BRN2 7 ** ENZ 30 6 CTR *** * − ** 5 +BRN2 +ENZ (10 µmol/L) *** 4 1.0 1.5 20 ** ** 1.0 * 0.5 10 ** *** * ** 0.5 ** Relative expression Relative expression 0 0.0 Relative proliferation 0.0 BRN2 NSE CGA SYP NCAM BRN2 AR PSA FK NK TMP CTROE BRN2

Figure 4. BRN2 expression inversely correlates with PSA in human prostate cancer and is induced by ENZ. A, BRN2 expression in CRPC tumors versus patient serum PSA (37). B, IHC score as well as representative TMA samples for BRN2 protein expression in radical prostatectomy (primary) or transurethral resection of the prostate (TURP) CRPC specimens from patients with circulating levels of PSA between 0 and 10 ng/mL or 40 and 50 ng/mL. Primary prostate cancer PSA 0–10 n = 26; primary prostate cancer PSA 40–50 n = 7; CRPC PSA 0–10 n = 22; CRPC PSA 40–50 n = 2. C, Protein and relative mRNA expression of BRN2, SYP, NSE, CGA, and VINC in siScr and siBRN2 16DCRPC cells treated ± 10 μmol/L ENZ for 2, 4, or 7 days. D, Left, Sanger sequencing result of CRISPR/ Cas9 introduction of POU3F2 deletion. The POU3F2-null clone #3 had monoallelic point mutations, whereas clone #7 had a biallelic 1-bp deletion leading to the frameshift mutation and stop codon. Mutated regions are labeled in green. D, Middle and right, protein and relative mRNA expression of BRN2, SYP, NSE, CGA, and VINC in different clones of 16DCRPC harboring BRN2 CRISPR knockout (BRN2 KO) ± 10 μmol/L ENZ for 7 days compared with wild-type 16DCRPC cells (WT). mRNA expression of BRN2 in 16DCRPC WT was set to 1. E, Relative mRNA expression of BRN2 and NE markers in 16DCRPC cells with overexpression of BRN2, treated with 10 μmol/L ENZ for 7 days compared with control vector treated (CRPC + ENZ = 1). F and G, Relative mRNA expression of AR and AR target genes in 16DCRPC cells overexpressing BRN2 and (G) growth response of 16DCRPC overexpressing BRN2 exposed to 10 μmol/L of ENZ after 72 hours. See also Supple- mentary Figs. S4 and S5. Statistics were performed on pooled data from at least three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

JANUARY 2017 CANCER DISCOVERY | OF8

RESEARCH ARTICLE Bishop et al.

AB C BRN2–Luc activity ENZR + R1881 AR–BRN2 ChIP CTR 2.0 0.15 1.25 +R1881 −R1881 + BRN2+R1881 **** +R1881 1.00 1.5 0.10 0.75 *** 1.0 0.50 * Input (%) 0.05 0.5 * * * * * Relative activity 0.25 * * *

Relative expression * 0.00 0.0 0.00 −R1881 +R1881 BRN2 NSE SYP CGANCAM IgGARE −CTR1 −CTR2 DE F BRN2 expression NE marker expression CRPC Ap R1881 vs. AREAp R1881 vs. AREAp 16D + ENZ vs. ARE CTR R1881 CTR 1.25 1.25 50 nmol/L+R1881 25 ENZ 100 nmol/L+R1881 AREAp 1.00 1.00 200 nmol/L+R1881 20 *** 0.75 0.75 * * 15 * * ** 0.50 0.50 * * * * * 10 ** * * * ** 0.25 *** *** * * ** * 0.25 * * * 5 * ** * *

Relative expression * Relative expression * * 0.00 Relative expression +− ++ + R1881 0.00 0 −−50 100 200 AREAp NSE SYP CGANCAM BRN2 NSE SYP CGANCAM (nmol/L)

Figure 5. BRN2 activity and BRN2-dependent neuroendocrine marker expression are suppressed by AR. A, Relative activity of BRN2–luciferase (Luc) reporter 48 hours after transfection in 42FENZR cells treated with 10 nmol/L R1881 for 24 hours compared with control-treated cells (= 1). Luciferase activity is normalized to Renilla. B, Relative mRNA expression of BRN2 and NE markers in CTR transfected and 42FENZR cells overexpressing BRN2 treated with 10 nmol/L R1881 for 24 hours compared with control-treated/untransfected cells (= 1). C, Chromatin immunoprecipitation (ChIP) showing AR binding to the enhancer region of BRN2 in 42FENZR cells treated ±10 nmol/L R1881 for 24 hours. D–F, Relative mRNA expression of (D) BRN2 and (E) NE markers in 42FENZR cells treated with 10 μmol/L R1881 ± increasing doses of BRN2–AREAp for 48 hours compared with control-treated cells (= 1) or (F) 16DCRPC cells treated with 10 nmol/L ENZ ± increasing doses of BRN2–AREAp for 7 days compared with control treated cells (= 1). See also Supplementary Fig. S6. Statistical analyses were performed on pooled data from at least three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Fig. S5G). Together, these results suggest that the suppres- lation with R1881 significantly increased AR occupancy sion of classic AR signaling with ENZ treatment leads to at this ARE compared with androgen-deprived conditions induction of an NE phenotype in CRPC that can be driven (Fig. 5C). To address the effect of AR binding specifically by the expression of the neural transcription factor BRN2, to BRN2 at this ARE, a 20-bp DNA aptamer, which physi- leading to ENZR. cally inhibits binding of other molecules to its complemen- tary sequence (39), was designed to prevent AR binding at AR Directly Represses BRN2 Expression the ARE in the BRN2 enhancer region, and the effects of and Activity R1881 on BRN2 and NE marker expression were assessed. To address AR regulation of BRN2 in CRPC, we assessed The reduction of BRN2 (Fig. 5D) and NE marker (Fig. 5E) R the effects of synthetic androgen (R1881) stimulation of expression by R1881 treatment in ENZ cells could be AR on BRN2 activity and expression. We found that R1881 rescued in a dose-dependent fashion by introduction of Ap significantly reduced BRN2 reporter activity in 42FENZR the BRN2 ARE aptamer (ARE ). No effects of aptamer (Fig. 5A), 42DENZR (Supplementary Fig. S6A), and in treatment were seen on other AR-dependent genes such as 16DCRPC cells treated with ENZ (Supplementary Fig. S6B). PSA, FKBP5, and TMPRSS2, indicating the specific effects of The expression of BRN2 mRNA was also reduced by R1881 aptamer treatment on AR binding to BRN2 (Supplemen- CRPC in multiple cell lines (Fig. 5B; Supplementary Fig. S6C– tary Fig. S6F). In addition, treatment of 16D cells with Ap S6E) and was accompanied by a reduction in terminal the BRN2 ARE increased expression of BRN2 and NE NE markers (Fig. 5B; Supplementary Fig. S6C and S6D). markers to similar levels as ENZ (Fig. 5F). Taken together, Importantly, the R1881-dependent reduced expression of these results indicate that BRN2 is negatively regulated by R CRPC NE markers was rescued by BRN2 overexpression, suggest- AR activation in both ENZ cells and 16D cells under ing that it is an important upstream androgen-regulated the pressure of ENZ. transcription factor responsible for NE marker expression (Fig. 5B; Supplementary Fig. S6C and S6D). Indeed, we BRN2 and AR Regulate SOX2 in identified an androgen response element (ARE) 8,733 bp NE Differentiation upstream of the BRN2 transcriptional start site, and chro- As in neural development, multiple transcription factors matin immunoprecipitation (ChIP) showed that stimu- may enhance an NE phenotype in prostate cancer. One

OF9 | CANCER DISCOVERY JANUARY 2017 www.aacrjournals.org

BRN2 Is an AR-Suppressed Driver of NEPC RESEARCH ARTICLE candidate that has been implicated in human NEPC (40) These results strongly implicated BRN2 over SOX2 in and cooperates with BRN2 in neural cells (41) is SOX2. driving an NE phenotype in CRPC cells. To investigate Indeed, SOX2 was significantly upregulated in human NEPC this hypothesis, we altered BRN2 and/or SOX2 expression compared with adenocarcinoma or CRPC (Fig. 6A) as well with siRNA or by overexpression in multiple prostate can- as in our NEPC PDX (Fig. 6B) and was positively correlated cer cell lines and found that although BRN2 knockdown with BRN2 expression in metastatic prostate cancer (ref. 37; reduced the expression of NE markers, SOX2 knockdown Supplementary Fig. S7A). Like BRN2, SOX2 expression was did not (Fig. 6J; Supplementary Fig. S7J). Reciprocally, highly expressed in NCIH660 cells, as well as in 42DENZR forced expression of SOX2 marginally increased expression and 42FENZR cells compared with other prostate cancer lines of terminal markers of NE differentiation in ENZ-treated (Fig. 6C). In addition, genes co-bound and co-regulated 16DCRPC and PC3 cells (Fig. 6K; Supplementary Fig. S7K), by both BRN2 and SOX2 identified in neural progenitor although not to the same extent as overexpression of BRN2 cells (NPC; ref. 42) were also enriched in patients with (Fig. 4E; Supplementary Fig. S5B), and did not increase NEPC (Fig. 6D, left) and in 42DENZR and 42FENZR compared NE markers in ENZ-treated LAPC4 cells (Supplementary with 16DCRPC cells (Fig. 6D, right; Supplementary Fig. S7B). Fig. S7L). Finally, in 16DCRPC and LAPC4 cells where we Coregulation of these genes may be mediated by BRN2– simultaneously silenced BRN2 and overexpressed SOX2, SOX2 protein–protein interaction, which we confirmed by we did not observe an increase in NE marker expression coimmunoprecipitation of BRN2 and SOX2 in 42DENZR and (Fig. 6L; Supplementary Fig. S7L), indicating that BRN2 is 42FENZR cells (Fig. 6E), and Re-ChIP experiments showing required for any SOX2-dependent induction of NE differ- BRN2 and SOX2 co-occupy enhancer regions of NES and entiation. Taken together, these data show, for the first RFX4 in 42DENZR (Fig. 6F). These results showed that BRN2 time, that BRN2-dependent control of SOX2 in prostate and SOX2 can physically interact in prostate cancer cells cancer is inhibited by the AR and drives the expression of and supported further investigation into the hypothesis that terminal NE markers. Moreover, they underscore the cen- these two transcription factors may work in concert to pro- tral role for BRN2 as a primary regulator of NEPC differ- mote NEPC. entiation in prostate cancer. Similar to BRN2, SOX2 is also negatively regulated by the AR (43). Consistently, we found that SOX2 expression BRN2 Is Required for NE Marker Expression R was induced by ENZ treatment in 16DCRPC and LAPC4 cells and Aggressive Growth of ENZ Cells In Vitro (Supplementary Fig. S7C), whereas R1881 reduced SOX2 and In Vivo expression (Supplementary Fig. S7D). Beyond the AR, how- Our data suggested that BRN2 contributes to ENZ ever, it remains unclear what regulates SOX2 expression in resistance and is a master regulator of ENZ-induced NE prostate cancer and whether BRN2 is involved. We found differentiation in CRPC. Confirming the requirement for that SOX2 mRNA levels were reduced after BRN2 knock- BRN2 in supporting an NE phenotype across multiple cell down in ENZR (Fig. 6G), PC3, NCIH660, and ENZ-treated lines, we found that transient targeting of BRN2 yielded a LAPC4 cells (Supplementary Fig. S7E–S7G), whereas over- marked reduction in mRNA levels of NSE, SYP, CGA, and expression of BRN2 in 16DCRPC (Fig. 6H), PC3, LAPC4, NCAM1 in NCIH660 (Fig. 7A), 42DENZR, 42FENZR, PC3, and or 49FENZR cells (Supplementary Fig. S7H) markedly LAPC4 cells (Supplementary Fig. S8A–S8D). Similar results increased SOX2, data that are in accordance with previous were observed in stable shBRN2 knockdown 42FENZR cells studies showing that BRN2 is required for SOX2 activity (Fig. 7B). In addition to the expression of terminal NE in neural development (44). Moreover, R1881-dependent markers, we questioned whether BRN2 may be important suppression of SOX2 could be rescued in ENZR cells by in regulating the aggressive biology of NEPC. Therefore, the addition of our BRN2 ARE aptamer (Supplemen- we investigated the effects of BRN2 knockdown on cellular tary Fig. S7I), further supporting our data that BRN2 is proliferation, migration, and invasion in vitro. BRN2 knock- AR-suppressed and regulates SOX2 expression. To inves- down significantly reduced proliferation in both siBRN2 tigate whether SOX2 may reciprocally regulate BRN2, NCIH660 cells and shBRN2 42FENZR cells (Fig. 7C and D) we examined the enhancer region of BRN2 for potential and prevented wound closure in a one-dimensional scratch SOX binding sites. Strikingly, we found a canonical SOX assay (Fig. 7E), as well as the capacity of shBRN2 42FENZR binding motif (42) that overlapped with the ARE in BRN2 cells compared with sh-control cells to migrate through a (Fig. 6I, diagram). ChIP was used to validate binding of Matrigel-coated Boyden chamber (Fig. 7F). Similar results SOX2 to the same region as the AR upstream of BRN2 in were observed in shBRN2 16DCRPC cells (Supplementary ENZR cells in androgen-deprived versus stimulated con- Fig. S8E–S8G). The reduced proliferative capacity of shBRN2 ditions, and we found that SOX2 was able to bind this site cells in vitro was translated in vivo; shBRN2 42FENZR subcuta- in the absence, but not in the presence, of R1881 (Fig. 6I). neous tumors grown in castrated mice under the pressure Interestingly, however, we found that neither knockdown of ENZ were smaller than sh-control tumors (Fig. 7G), and nor overexpression of SOX2 altered mRNA expression these tumors had reduced expression of BRN2 and terminal of BRN2 in ENZ-treated 16DCRPC cells (Fig. 6J and K) or NE markers (Fig. 7H), indicating that this NEPC signature NCIH660-, PC3-, or ENZ-treated LAPC4 cells (Supplemen- was associated with more aggressive growth. These in vitro tary Fig. S7J–S7L). These results suggest a unidirectional and in vivo results indicate that ENZ-induced NE differ- regulation of SOX2 by BRN2 and not vice versa, and show entiation is mediated by BRN2, which can be targeted to that SOX2 binding to the BRN2 enhancer is regulated reduce invasiveness and tumor proliferation in both ENZR by AR. and CRPC.

JANUARY 2017 CANCER DISCOVERY | OF10

RESEARCH ARTICLE Bishop et al.

AB C SOX2 expression SOX2 expression SOX2 expression in PDX human PCa in PCa cell lines **** 10,000 3,000 **** 2,500 40 1,000 2,000 25 *** *** 100 20 10 20 15 ** ***

read counts 10 5 * ** Log2 (RPKM + 1) Relative expression 0 Normalized RNA-seq 1 0 Adeno CRPC NEPC Adeno NEPC PC3 C4-2 VCaP LNCaPCRPC LAPC4DU145 42DENZR42FENZRNCIH660 D EF SOX2–BRN2 NPC targets BRN2/SOX2 Co-occupancy 0.10 LRRN1 LRRN1 27x IgG NES NES SOX2/BRN2 IP: BRN2 0.08 *** SOX2 HDAC9 BRN2/SOX2 HDAC9 SOX2OT Inp 16D 42D 42F SOX2 0.06 RFX4 RFX4 BRN2 * ID4 ID4 Input (%) 0.04 ** KIRREL3 OCT4 COPS2 KIRREL3 SOX2 0.02 *** OCT4 OCPS2 1x Adeno CRPC NE Adeno NE 16D 42D 0.00 CRPC ENZR NES NES RFX4 RFX4 Beltran, 2016 Beltran, 2011 neg neg G HI TTTCTGTT BRN2 SOX2 expression: SOX2 expression: AACAATTTTTCT siBRN2 BRN2 overexpression 0.10 **** −R1881 CTR 0.08 +R1881 1.00 CTR 45 OE BRN2 siBRN2 0.75 35 *** 0.06

0.50 25 0.04

* ** Input (%) 15 0.25 ** 0.02 5 Relative expression Relative expression 0.00 0.00 BRN2 SOX2 BRN2 SOX2 IgG ARE Neg1 Neg2 JKL siBRN2 vs. siSOX2 SOX2 overexpression SOX2 overexpression/siBRN2

2.5 CTR 100 *** 6.0 *** CTR CTR siSOX2 80 5.0 +SOX2 +SOX2/siBRN2 2.0 siBRN2 60 4.0 5 1.5 4 * 1.0 * 1.0 3 ** * * * * * 2 0.5 0.5 * * * * * * * * * 1

* Relative expression Relative expression 0.0 Relative expression 0 0.0 BRN2 SOX2 NSE SYP CGANCAM BRN2 SOX2 NSE SYP CGA NCAM BRN2 SOX2 NSE SYP CGA NCAM

Figure 6. BRN2-dependent regulation of SOX2 expression is inhibited by the AR and drives NE differentiation. A, SOX2 expression in human prostatic adenocarcinoma (Adeno, n = 68), CRPC (n = 32), and NEPC tumors (n = 21; refs. 11, 33; line at mean). B, SOX2 expression in patient-derived prostatic adenocarcinoma xenografts (Adeno) and terminally transdifferentiated NEPC tumors (NEPC) based on RNA sequencing (RNA-seq). C, Relative mRNA expression of SOX2 across different prostate cancer cell lines compared with LNCaP as control (= 1). D, Left, heat map of SOX2–BRN2 co-bound neural progenitor cell (NPC) gene targets (42) in NEPC, Adeno, and CRPC tumors in two cohorts (11, 33). D, Right, heat map of fold increase in reads per million of genes identified as co-bound by SOX2 and BRN2 in NPCs (42) in 42DENZR cells compared with 16DCRPC (= 1). E, SOX2–BRN2 protein interaction shown by immunoprecipitation of BRN2 and Western blot for SOX2 in 16DCRPC, 42DENZR, and 42FENZR cells. F, Quantitative Re-ChIP analysis for the enhancer region of NES and RFX4 in 42DENZR cells using anti-BRN2 and anti-SOX2 antibodies with indicated sequential order (BRN2/SOX2, blue; or SOX2/BRN2, red). IgG was used as antibody control, and sequences outside of the enhancer regions were designed for the specificity of the binding. G, Relative mRNA expression of BRN2 and SOX2 in 42FENZR transfected with siBRN2 compared with siCTR transfected cells (= 1). H, Relative mRNA expression of BRN2 and SOX2 in 16DCRPC cells overexpressing BRN2 (OE BRN2) compared with control vector (CTR = 1). I, Chromatin immunoprecipitation showing SOX2 binding to the enhancer region of BRN2 in 42FENZR cells treated with 10 nmol/L R1881 for 24 hours compared with control-treated cells (= 1). J–L, Relative mRNA expression of BRN2, SOX2, and NE markers in 16DCRPC cells transfected with (J) siSOX2, siBRN2, or siCTR; K, SOX2 overexpression vector (OE SOX2) or CTR vector; L, SOX2 overexpression vector + siBRN2 (OE SOX2/siBRN2) or CTR vector. Post-transfection, 16DCRPC cells were cultured in the presence of 10 μmol/L ENZ and harvested 5 days later for analysis. For each panel, relative mRNA values are compared with CTR = 1. See also Supplementary Fig. S7. Statistical analyses were performed on pooled data from at least three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. PCa, prostate cancer.

OF11 | CANCER DISCOVERY JANUARY 2017 www.aacrjournals.org

BRN2 Is an AR-Suppressed Driver of NEPC RESEARCH ARTICLE

A B NCIH660 siBRN2 NE markers 1.00 1.00

0.75 0.75 * ENZR * 42F shCTR 0.50 * ** 0.50 ** *** 42FENZR shBRN2 0.25 *** 0.25 *** Relative expression Relative expression 0.00 0.00 ND ND BRN2 NSE SYP CGA NCAM BRN2 NSE SYP CGA NCAM siCTR siBRN2

CD E NCIH660 proliferation Proliferation Migration 1.5 siCTR 2.0 8 siBRN2 6 1.0 1.5 4 ** 1.0 0.5 2 *** 0.5 0 ABS 590 nmol/L Relative proliferation 0.0 0.0 −2 siCTR siBRN2 42FENZR 42FENZR Relative wound density (%) 0612 18 24 shCTR shBRN2 Time (hours)

FG H Invasion Tumor volume Tumor NE markers 60 2,000 1.5 ) 3 1,500 40 1.0 ** 1,000 20 0.5 ** 500 ** # Invaded cells ** mor volume (mm Tu 0 0 Relative expression 0.0 42FENZR 42FENZR 024681012 BRN2 NSE SYP CGA NCAM shCTR shBRN2 Time (weeks)

Figure 7. BRN2 is required for neuroendocrine marker expression and aggressive growth of ENZR cells in vitro and in vivo. A and B, Relative mRNA expression of BRN2 and NE markers in (A) NCIH660 cells transfected with BRN2 siRNA (siBRN2) compared with control (siCTR = 1) and (B) 42FENZR cells with stable BRN2 knockdown (shBRN2) compared with control transfected cells (shCTR =1). C, Relative proliferation, 72 hours after seeding in NCIH660 cells transfected with BRN2 siRNA (siBRN2) compared with control (siCTR = 1). D–F, Relative proliferation (D), relative wound density in one-dimensional scratch assay (E) and number of cells migrated through Matrigel Boyden chamber (F) in 42FENZR cells with stable BRN2 knockdown (shBRN2) compared with control-transfected cells (shCTR =1). G, Tumor volume of 42FENZR shCTR and 42FENZR shBRN2 xenografts grown in vivo (n = 10). H, Relative mRNA expression of BRN2 and NE markers in 42FENZR shBRN2 versus shCTR xenografts (= 1) harvested at 12 weeks after inoculation. Graph represents pooled data from 6 shBRN2 and 6 shCTR tumors. See also Supplementary Fig. S8. Statistical analyses were performed on pooled data from at least three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

DISCUSSION pressed by the AR that is both sufficient and required for NE differentiation in prostate cancer and mediates resistance to Clinical evidence suggests that the repercussion of potent ENZ. Secondly, our data show for the first time the BRN2- AR suppression with APIs for a subset of patients with dependent regulation of SOX2 in prostate cancer, and the CRPC is the development of highly lethal NEPC tumors importance of BRN2 over SOX2 in promoting NEPC. Lastly, (20). Our work reveals several novel findings with impli- our data reveal a striking overlap of AR and SOX binding cations for patients with CRPC and NEPC. First and fore- motifs in the enhancer region of BRN2 that allow the AR most, we identify the master neural transcription factor to competitively inhibit the interaction between SOX2 and BRN2 as a central and clinically relevant driver of NE the BRN2 enhancer. These in vitro, in vivo, and human stud- marker expression in advanced prostate cancer. Utilizing a ies highlight BRN2 as a key driver of NE differentiation “non–AR-driven” in vivo–derived model of ENZR- and ENZ- that may indicate progression toward a non–AR-driven or treated CRPC, we identified BRN2 as a direct target sup- NE phenotype in patients with prostate cancer. Moreover,

JANUARY 2017 CANCER DISCOVERY | OF12

RESEARCH ARTICLE Bishop et al. they suggest that relieving AR suppression of BRN2 could with NEPC (40), and is required for the function and main- be a central mechanism driving NE differentiation, making tenance of NPCs (54). Importantly, however, SOX2 can BRN2 a strong potential therapeutic target for the treat- only drive a neural development program by cooperating ment and/or prevention of NEPC. with other master transcription factors, especially POU BRN2 is a POU-domain transcription factor well described family members (41). In particular, BRN2 and SOX2 co- in developmental biology, where it plays an essential role in bind upstream of many genes with central roles in neural neural cell differentiation (32). In addition, BRN2 is highly cell fate and function (42, 55). Moreover, BRN2 is a key expressed in NE small cell lung cancer (SCLC), where it acts regulator of SOX2 activity in NPCs, a function that is highly upstream of other key regulators of neural programming (45) evolutionarily conserved (44, 55). Despite extensive research and is required for aggressive tumor growth (46). Comple- into this “pou-sox code” in neural development, how SOX2 menting these reports, BRN2 was a highly expressed master is regulated in prostate cancer is largely unexplored. Our neural transcription factor identified by RNA sequencing of study shows that BRN2 is required for SOX2 expression our non–AR-driven ENZR cell lines, making it our top candi- in both CRPC cells treated with ENZ as well as ENZR cells. date for a potential driver of ENZ-induced NE differentiation The regulation of SOX2 by AR may be compounded by in prostate cancer. Indeed, we found that BRN2 was not only direct enhancer binding (43) and via AR-dependent sup- sufficient to increase terminal markers of NE differentiation pression of BRN2. Indeed, as we found for BRN2, SOX2 in CRPC but was also required for their expression in CRPC is more highly expressed in AR- than AR+ prostate cancer cells exposed to ENZ, in ENZR cell lines and in bona fide NEPC cell lines (40) and is higher in NEPC and metastatic CRPC NCIH660 cells. Importantly, inhibiting BRN2 expression than in adenocarcinoma (40, 43, 56). Altogether, these data and concomitant reduction in NE markers in both ENZR suggest that the link between SOX2 and progression to AR- and CRPC cells functionally led to significantly reduced independent CRPC or NEPC may be a result of increased proliferation, migration, and invasion, as well as decreased BRN2 expression. ENZR tumor growth in vivo and increased resistance to ENZ Our results showing BRN2-mediated regulation of SOX2 in vitro. These data mirror other reports showing the require- also shed light on the importance of BRN2 over SOX2 in ment for BRN2 in SCLC (46, 47) as well as in melanoma, driving NE differentiation in prostate cancer cells. Although where it is required for invasion and migration (48–50). SOX2 is present in NE tumors of not only the prostate (40) Importantly, we have identified BRN2 as a major regulator but also the lung (57, 58) and skin (59), these studies have not of NE differentiation in a model of castration- and ENZ- shown a direct requirement of SOX2 in supporting this phe- resistant prostate cancer and in terminally differentiated notype. Our results suggest that whereas SOX2 overexpres- NEPC cell lines. sion alone in CRPC cells can marginally increase expression Data from human specimens brought clinical relevance to of CGA, NSE, SYP, and NCAM1, SOX2 requires the presence BRN2 in aggressive prostate cancer tumors, including NEPC. of BRN2 to significantly upregulate the expression of these RNA-sequencing data from patient cohorts showed that NE markers. The hypothesis that BRN2 and SOX2 work BRN2 was most highly expressed in clinically defined NEPC together to drive a neural program in CRPC cells is further tumors compared with adenocarcinomas. Importantly, how- supported by our data showing direct BRN2–SOX2 pro- ever, BRN2 was identified in CRPC specimens as well, indicat- tein–protein interaction at the enhancer region of neuronal ing that BRN2 is not a specific marker of NEPC, but rather genes, leading to their upregulation in ENZR cells, which is in may indicate potential toward disease progression, especially accordance with previously reported ChIP-seq analysis (42). in a setting of androgen deprivation. This was in accordance Importantly, however, although BRN2 and SOX2 can both with our data showing that BRN2 is inducible in CRPC cells bind upstream of each other in NPCs (42), the controlling under the pressure of ENZ and supported our hypothesis signals as well as the consequences of these binding events that it is an androgen-suppressed gene. Indeed, we found remain unclear. Intriguingly, our observation that SOX2 BRN2 was most highly expressed in primary adenocarcinoma bound to a canonical SOX motif that overlapped with the or CRPC tissue from patients with low levels of circulating ARE in the enhancer region of BRN2 suggests that the AR PSA, and BRN2 expression significantly increased from pro- may play an important role in controlling SOX protein DNA gression to CRPC only in patients with low PSA levels. Pub- binding in prostate cancer cells. Although the consequence licly available data mirrored this inverse correlation between of this inhibition of SOX2 binding to the BRN2 enhancer in high BRN2 expression and low circulating PSA, and further the presence of androgen remains unclear, it may be that this underscored the association between BRN2 and the potential interaction supports the two factors coordinating to drive for NE-like disease, as it positively correlated with SYP and downstream neural gene expression. CGA expression. Although multiple pathways likely converge to drive the Human data implicating AR control of BRN2 were com- emergence of NEPC, understanding the contribution of the plemented by mechanistic studies using ENZR cell lines, AR to this disease is critical for better implementation of which showed that BRN2 is suppressed by the AR through current API therapies and novel drug design. Together, data ligand-dependent binding to an ARE in the enhancer region from our human cohorts and in vitro mechanistic analy- of BRN2. AR binding to this site resulted in reduced lev- sis strongly implicate BRN2 as an androgen-suppressed els of not only NE markers but also SOX2, which we also transcription factor that plays a significant role in the pro- found highly expressed in human NEPC. SOX2 is a well- gression of prostate cancer from adenocarcinoma to NEPC, defined transcription factor that supports the proliferation making it a potentially attractive and novel therapeutic and invasiveness of prostate cancer (43, 51–53), is associated target.

OF13 | CANCER DISCOVERY JANUARY 2017 www.aacrjournals.org

BRN2 Is an AR-Suppressed Driver of NEPC RESEARCH ARTICLE

METHODS ences) 18 to 24 hours prior to transfection with siRNA, shRNA, plasmid overexpression, or DNA aptamer. Generation of ENZR Xenografts and Cell Lines The detailed procedure for generation of CRPC and ENZR tumors siRNA and cell lines is found in Supplementary Methods as well as our Cells were transfected with 10 nmol/L BRN2#1 or control siRNA previously published report (60). A schematic depicting model gen- (Santa Cruz Biotechnology), 10 nmol/L of BRN2#2 and siSOX2 (Life eration and growth of individual xenografts from which cell lines are Technology) using Oligofectamine (Invitrogen), and OPTI-MEM derived is shown in Supplementary Fig. S1. media (Gibco). After 18 hours, cells were retransfected. After 4 hours, OPTI-MEM media were replaced with complete media, and cells were Cell Line Culture and Reagents harvested after 48 hours. PC3, NCIH660, and LAPC4 cells were obtained from the ATCC in 2013. LNCaP cells were kindly provided by Dr. Leland W.K. Chung shRNA (Emory University) and authenticated in January 2013. CRPC and The same protocol as siRNA was used for shBRN2 transfections R ENZ cell lines were generated from LNCaP cells (60), tested, and using shBRN2 or control shRNA (Santa Cruz Biotechnology) and authenticated by whole-genome and whole-transcriptome sequenc- successfully transfected clones were selected for and expanded in ing (Illumina Genome Analyzer IIx, 2012). Cells were maintained in complete media containing 10 μg/mL puromycin. RPMI-1640 (LNCaP-derived and PC3) or Iscove’s Modified Dulbec- co’s Medium (LAPC4), containing 10% FBS, 100 U/mL penicillin-G, Overexpression 100 mg/mL streptomycin (all Hyclone), ±10 μmol/L ENZ (Haoyuan SOX2 plasmid (1 μg; Addgene, #16353) or 8 μg BRN2 plasmid Chemexpress), or DMSO (Sigma-Aldrich) vehicle (for ENZR vs. (Addgene, #19711) was transfected using Mirus T20/20 and OPTI- CRPC, PC3 did not receive ENZ). Where indicated, CRPC or LAPC4 MEM media (Invitrogen) according to the manufacturer’s instruc- cells induced to an NE phenotype were cultured in RPMI-1640, 10% tions. After 18 to 24 hours, OPTI-MEM media were replaced with FBS, 100 U/mL penicillin-G, 100 mg/mL streptomycin, +10 μmol/L complete media 10 μmol/L ENZ, and cells were harvested after 7 ENZ for 7 days prior to downstream analysis. Cells were seeded at a ± days or at indicated time points. For 7-day experiments, CRPC cells density of 106 cells/10 mL media and harvested after 72 hours unless were retransfected on day 4. otherwise noted.

Generation of BRN2 CRISPR Knockout Cells siRNA/Overexpression Cells were plasmid transfected with Mirus T20/20 on day 1 and Cells were transfected with 500-ng GeneArt Platinum Cas9 nucle- the following day transfected with siRNA with Oligofectamine. After ase (Thermo) and 125-ng guide RNA (gRNA) using Lipofectamine 18 to 24 hours cells were retransfected with siRNA, and OPTI-MEM CRISPRMAX (Thermo). The targeting gRNA sequence 5′-GCTG- was replaced with complete media ±10 μmol/L ENZ and harvested TAGTGGTTAGACGCTG-3′ was used to edit exon 1 of the POU3F1 48 hours later (total 5 days). locus. At 72 hours after transfection, cells were harvested for analysis of genome modification efficiency using the GeneArt Genomic Cleavage Detection Kit (Thermo) with the forward primer 5′-AAAT- BRN2 Aptamer CAAAGGGCGCGGCGCC-3′ and reverse primer 5′-GCCGCCGC- Indicated doses of BRN2–ARE aptamer were transfected into cells CGTGGGACAG-3′. Ten individual clones were isolated and assessed using Oligofectamine. After 18 hours, cells were transfected for a sec- for indels at the POU3F2 locus by Sanger sequencing. ond time for 4 hours. This was repeated on day 4, and cells were either in OPTI-MEM or in complete media + 10 μmol/L ENZ. 42DENZR Cell Line and Tumor Microarray and RNA Sequencing cells were transfected with the aptamer, and after 18 hours transfected a second time. Following this, they were exposed to R1881 for Microarray gene expression was performed as previously described 48 hours, and samples were harvested. For aptamer sequence please (36) using Agilent SurePrint G3 Human GE 8 × 60 K slides (Design see Supplementary Table S4. ID 028004) and analyzed using Agilent GeneSpring 11.5.1 and Inge- nuity Knowledge Base (Ingenuity Systems). Specimens were prepared qRT-PCR for RNA sequencing using the TruSeq RNA Library Preparation Kit v2, and transcriptome sequencing (RNA-seq) was performed using Total RNA was extracted from cells using TRIzol reagent (Life Illumina HiSeq 2000 (Illumina Inc.) or HiSeq 2500 (human tumors) Technology) and 2 μg was reversed transcribed using MMLV reverse according to standard protocols. Sequence data mapping and pro- transcriptase and random hexamers (Invitrogen). Real-time PCR cessing was performed as previously described (61), except nor- was performed using SyberGreen ROX Master Mix (Roche Applied malization was performed using reads per million. Quantification Science). Target gene expression was normalized to GAPDH levels of gene expression was performed via RSEQtools using GENCODE in three experimental replicates per sample. For primer sequences, v19 as reference gene annotation set. Expression levels (RPKM) were please see Supplementary Table S4. estimated by counting all nucleotides mapped to the gene and normalized by the total number of mapped nucleotides (per million) Immunoprecipitation and Western Blotting and the gene length (per kilobase). Sequencing of the AR ligand Immunoprecipitation was performed using the ImmunoCruz IP/ binding domain was performed exactly as previously described (26). WB Optima B System (Santa Cruz Biotechnology) based on the To assess global differences in gene expression in microarray and manufacturer’s guideline. Dilution (1/50) of primary antibody was RNA-sequencing data, multidimensional scaling to analyze differ- used for immunoprecipitation. Total proteins were extracted from ences between cell line gene expression data was performed using the adherent cells grown in vitro using RIPA lysis buffer. Forty micro- PCoA tool in XLSTAT software (Addinsoft). grams of protein was resolved by SDS-PAGE, and the following antibodies were used for Western blot: AR, PSA, BRN2, SYP (Cell Cell Line Transfection Signaling Technology), NSE (Dako), SOX2 (Millipore), and CGA CRPC or ENZR cells were seeded at a density of 106 cells/10 mL and vinculin (Sigma-Aldrich). Blots were incubated overnight at 4°C complete media in 10-cm tissue culture dishes (Corning Life Sci- with designated primary antibodies at 1:1,000 dilution, unless noted

JANUARY 2017 CANCER DISCOVERY | OF14

RESEARCH ARTICLE Bishop et al. otherwise. Proteins were visualized using the Odyssey system (Li-Cor stream of the ARE 8733 bp upstream of BRN2 TSS using Primer Biosciences). Express 3 (Thermo Fisher). Re-ChIP was performed via a Re-ChIP- IT kit (Active Motif Inc.), based on the manufacturer’s protocol Luciferase Assay using antibodies against BRN2 (Cell Signaling Technology) and CRPC and ENZR cells were plated in 6-well plates (2 × 104 cells/ SOX2 (Millipore). IgG was used as a negative control for anti- cm2) and transfected with a BRN2 luciferase reporter (courtesy of bodies, and negative control primers for each binding site were Dr. Goding, Ludwig Institute for Cancer Research, Oxford, UK; ref. designed as listed in Supplemental Experimental Procedures. Using 62) or Renilla (Promega) using Lipofectin (Invitrogen). The total Primer Express 3 (Thermo Fisher), primers were designed around plasmid DNA used was normalized to 0.5 μg per well by the addi- BRN2/SOX2 consensus binding elements approximately 7,500 bp tion of Renilla. At 24 hours after transfection, cells were incubated upstream of NES start codon and 35,000 bp downstream of RFX4 with or without 10 nmol/L R1881 for 24 hours, and luciferase activi- start codon. Negative control primers were approximately 5,400 ties were measured using a Dual-Luciferase Reporter Assay System bp upstream of NES start codon and 38,000 bp downstream of (Promega) and a Tecan Infinite 200 PRO microplate luminometer RFX4 start codon. For primer sequences please see Supplementary (Tecan). BRN2 luciferase activities were corrected by the correspond- Table S4. ing Renilla luciferase activities. Results are expressed in arbitrary light units. Xenograft Studies shBRN2 and shCTR ENZR tumors were grown and monitored in Flow Cytometry vivo in castrated male athymic mice (Harlan Sprague-Dawley) under Cells were harvested using citric saline for 10 minutes at room tem- pressure of daily oral 10 mg/kg ENZ. Tumors were monitored for perature and washed in RPMI+10% FBS. Before antibody addition, growth, and blood was drawn for PSA screening weekly as previ- 3 cells were incubated with human Fc receptor Binding Inhibitor ously described (60). When tumors reached 1,000 mm or greater (eBioscience) for 20 minutes on ice. Flow cytometry staining was than 10% animal body weight, tumors were harvested and pro- performed using anti-human NCAM1 followed by staining with cessed for RNA analysis by qRT-PCR. All animal procedures were viability marker 7-aminoactinomycin D (7-AAD; both eBioscience, conducted according to the guidelines of the Canadian Council on per instructions) and fixation in 2% paraformaldehyde (PFA). Data Animal Care. were acquired (minimum 30K events) on a Canto II (BD Biosciences) and analyzed with FlowJo (TreeStar). Human Prostate Cancer Specimens for RNA-seq and Immunohistochemistry Proliferation Assay RNA-seq was performed as above on samples from the Weill The WST-1 (Promega) assay was used to assess cell growth accord- Cornell College of Medicine: Beltran 2016 (phs000909.v.p1, ing to the manufacturer’s protocol. One thousand cells were seeded cBioportal; ref. 33) = 68 adenocarcinoma, 34 CRPC-Adeno, and in 96-well plates in complete media, and absorbance at 450 nm was 15 CRPC-NE; Beltran 2011 (phs000310 .v.p1, cBioportal) = 30 measured after 72 hours. Adeno and 7 NEPC (11). Tumors were classified by the following criteria based on histomorphology (11, 33): Adeno, usual pros- Migration Assay tate adenocarcinoma without neuroendocrine differentiation (from radical prostatectomy); CRPC, tumor obtained from CRPC Cell migration was assessed in a wound-healing assay. Cells were adenocarcinoma metastasis without neuroendocrine differen- plated on Essen ImageLock 96-well plates (Essen Instruments) and tiation; and NEPC, either of the following categories, adenocar- incubated for 2 hours with mitomycin (5 μg/mL) prior to wound cinoma with >20%, neuroendocrine differentiation, small cell scratching with a wound scratcher (Essen Instruments) 24 hours carcinoma, large cell neuroendocrine carcinoma, or mixed small after plating. Wound confluence was monitored with the IncuCyte cell carcinoma—adenocarcinoma. For IHC, prostate cancer speci- Zoom Live-Cell Imaging System and software (Essen Instruments). mens were obtained from the Vancouver Prostate Centre Tissue Wound closure was measured every 6 hours for 24 hours by compar- Bank and were classified as above (Adeno, n = 93; CRPC, n = ing the mean relative wound density of three replicates. 30; NEPC, n = 11). Tissue microarrays of duplicate 1-mm cores were constructed manually (Beecher Instruments). Samples were Invasion Assay from radical prostatectomy or transurethral resection of pros- Invasion was assessed by the invasion of 2.5 × 104 cells through tate. Immunohistochemical staining was conducted as previously BioCoat Matrigel-coated Transwell inserts with 8-μm pore size (BD described (36) using the Ventana DiscoverXT Autostainer (Ven- Biosciences). After 24 hours, the Transwell insert was removed and tana Medical System) with enzyme labeled biotin streptavidin fixed for 10 minutes in 100% methanol (Sigma-Aldrich) at −20°C system and a solvent-resistant DAB Map Kit by using 1 of 150 and mounted on glass cover slips with Vectashield Mounting Media concentration of BRN2 (Abcam) and 1 of 25 concentrations of with DAPI (Vector Laboratories). Filters were imaged using Zeiss CGA and AR (Sigma) antibodies. Specimens were graded from Axioplan II microscope (Zeiss) and cells invaded in membrane were 0 to +3 intensity by visual scoring, representing negative-heavy quantified. staining. Automated quantitative image analysis was conducted using pro-plus image software. Scoring was conducted at 200× ChIP and Sequential ChIP (Re-ChIP) magnification. Cells treated with or without 1 nmol/L R1881 for 24 hours were cross-linked with PFA (Sigma-Aldrich) and sonicated to shear Statistical Analysis and Data Representation DNA. ChIP assay was performed using the ChIP Assay Kit (Aga- Pearson correlations (95% confidence interval) were performed rose Beads) according to the manufacturer’s protocol (Millipore) using GraphPad Prism (GraphPad Software) for IHC scoring and and antibodies against AR (N20; Santa Cruz Biotechnology) and BRN2 versus serum PSA. In bar graphs, unpaired, two-tailed, SOX2 (Millipore). Negative control primers were designed for the Student t tests were performed to analyze statistical significance regions approximately 1600 bp upstream and 1800 bp down- between groups using GraphPad Prism (GraphPad Software).

OF15 | CANCER DISCOVERY JANUARY 2017 www.aacrjournals.org

BRN2 Is an AR-Suppressed Driver of NEPC RESEARCH ARTICLE

Significance is indicated as follows: *,P < 0.05; **, P < 0.01; ***, 3. Beltran H, Tomlins S, Aparicio A, Arora V, Rickman D, Ayala G, et al. P < 0.001; ****, P < 0.0001. Graphs show pooled data with error Aggressive variants of castration-resistant prostate cancer. Clin Can- bars representing SEM obtained from at least three independent cer Res 2014;20:2846–50. experiments. 4. Small EJ, Huang J, Youngren J, Sokolov A, Aggarwal RR, Thomas G, et al. Characterization of neuroendocrine prostate cancer (NEPC) in patients with metastatic castration resistant prostate cancer Disclosure of Potential Conflicts of Interest (mCRPC) resistant to abiraterone (Abi) or enzalutamide (Enz): Pre- C.C. Collins is a consultant/advisory board member for Accuragen. liminary results from the SU2C/PCF/AACR West Coast Prostate Can- No potential conflicts of interest were disclosed by the authors. cer Dream Team (WCDT). ASCO Annual Meeting 2015;33 (abstract 5003). Authors’ Contributions 5. Epstein JI, Amin MB, Beltran H, Lotan TL, Mosquera JM, Reuter VE, et al. Proposed morphologic classification of prostate cancer Conception and design: J.L. Bishop, D. Thaper, S. Vahid, M.E. Gleave, with neuroendocrine differentiation. Am J Surg Pathol 2014;38: A. Zoubeidi 756–67. Development of methodology: J.L. Bishop, D. Thaper, S. Vahid, 6. Aparicio A, Logothetis CJ, Maity SN. Understanding the lethal vari- A. Davies, K.M. Nip, A. Zoubeidi ant of prostate cancer: power of examining extremes. Cancer Discov Acquisition of data (provided animals, acquired and managed 2011;1:466–8. patients, provided facilities, etc.): J.L. Bishop, D. Thaper, S. Vahid, 7. Alanee S, Moore A, Nutt M, Holland B, Dynda D, El-Zawahry A, et al. H. Kuruma, A. Angeles, F. Johnson, M.E. Gleave, M.A. Rubin, Y. Wang Contemporary incidence and mortality rates of neuroendocrine pros- Analysis and interpretation of data (e.g., statistical analysis, bio- tate cancer. Anticancer Res 2015;35:4145–50. statistics, computational analysis): J.L. Bishop, D. Thaper, S. Vahid, 8. Tan HL, Sood A, Rahimi HA, Wang W, Gupta N, Hicks J, et al. Rb loss K. Ketola, H. Kuruma, R. Jama, K.M. Nip, A.W. Wyatt, C.C. Collins, is characteristic of prostatic small cell neuroendocrine carcinoma. Y. Wang, H. Beltran, A. Zoubeidi Clin Cancer Res 2014;20:890–903. 9. Tzelepi V, Zhang J, Lu JF, Kleb B, Wu G, Wan X, et al. Modeling a Writing, review, and/or revision of the manuscript: J.L. Bishop, lethal prostate cancer variant with small-cell carcinoma features. Clin D. Thaper, S. Vahid, A. Davies, K. Ketola, R. Jama, C.C. Collins, Cancer Res 2012;18:666–77. H. Beltran, A. Zoubeidi 10. Chen H, Sun Y, Wu C, Magyar CE, Li X, Cheng L, et al. Pathogenesis Administrative, technical, or material support (i.e., reporting of prostatic small cell carcinoma involves the inactivation of the P53 or organizing data, constructing databases): J.L. Bishop, R. Jama, pathway. Endocr Relat Cancer 2012;19:321–31. K.M. Nip, A. Angeles, D. Lin, C.C. Collins 11. Beltran H, Rickman DS, Park K, Chae SS, Sboner A, MacDonald Study supervision: J.L. Bishop, A. Zoubeidi TY, et al. Molecular characterization of neuroendocrine prostate Other (pathology): L. Fazli cancer and identification of new drug targets. Cancer Discov 2011;1: 487–95. Acknowledgments 12. Akamatsu S, Wyatt AW, Lin D, Lysakowski S, Zhang F, Kim S, et al. The placental gene PEG10 promotes progression of neuroendocrine Tissue procurement and tissue microarray construction at the prostate cancer. Cell Rep 2015;12:922–36. Vancouver Prostate Centre were supported by the Terry Fox New 13. Lapuk AV, Wu C, Wyatt AW, McPherson A, McConeghy BJ, Brahmb- Frontiers Program Project Grant. We thank Dr. Goding (Ludwig hatt S, et al. From sequence to molecular pathology, and a mech- Institute for Cancer Research, Oxford, UK) for providing the BRN2 anism driving the neuroendocrine phenotype in prostate cancer. luciferase plasmid and Mary Bowden and Darrell Trendall for techni- J Pathol 2012;227:286–97. cal support in animal studies. 14. Zhang X, Coleman IM, Brown LG, True LD, Kollath L, Lucas JM, et al. SRRM4 expression and the loss of REST activity may promote the Grant Support emergence of the neuroendocrine phenotype in Castration-Resistant Prostate Cancer. Clin Cancer Res 2015;21:4698–708. This work is supported by the Prostate Cancer Foundation USA 15. Svensson C, Ceder J, Iglesias-Gato D, Chuan YC, Pang ST, Bjartell A, and Prostate Cancer Canada, and proudly funded by the Movember et al. REST mediates androgen receptor actions on gene repression Foundation (T2013-01). J.L. Bishop was supported by the Prostate and predicts early recurrence of prostate cancer. Nucleic Acids Res Cancer Foundation Young Investigator Award and the Urology Care 2014;42:999–1015. Foundation. K. Ketola was supported by the U.S. Department of 16. Wee ZN, Li Z, Lee PL, Lee ST, Lim YP, Yu Q. EZH2-mediated inactiva- Defense (PC121341). D. Thaper and A. Davies were supported by a tion of IFN-gamma-JAK-STAT1 signaling is an effective therapeutic Canadian Institute of Health Research scholarship. target in MYC-driven prostate cancer. Cell Rep 2014;8:204–16. The costs of publication of this article were defrayed in part by 17. Li Y, Donmez N, Sahinalp C, Xie N, Wang Y, Xue H, et al. SRRM4 drives the payment of page charges. This article must therefore be hereby neuroendocrine transdifferentiation of prostate adenocarcinoma marked advertisement in accordance with 18 U.S.C. Section 1734 under androgen receptor pathway inhibition. Eur Urol 2016 May solely to indicate this fact. 11. pii: S0302-2838(16)30138-5. doi: 10.1016/j.eururo.2016.04.028. [Epub ahead of print]. Received October 22, 2015; revised October 20, 2016; accepted 18. Abrahamsson PA. Neuroendocrine differentiation and hormone- October 21, 2016; published OnlineFirst October 26, 2016. refractory prostate cancer. Prostate Suppl 1996;6:3–8. 19. Burchardt T, Burchardt M, Chen MW, Cao Y, de la Taille A, Shabsigh A, et al. Transdifferentiation of prostate cancer cells to a neuroendocrine cell phenotype in vitro and in vivo. J Urol 1999;162:1800–5. REFERENCES 20. Hu CD, Choo R, Huang J. Neuroendocrine differentiation in prostate 1. Ferraldeschi R, Welti J, Luo J, Attard G, de Bono JS. Targeting the cancer: a mechanism of radioresistance and treatment failure. Front androgen receptor pathway in castration-resistant prostate cancer: Oncol 2015;5:90. progresses and prospects. Oncogene 2015;34:1745–57. 21. Kani K, Malihi PD, Jiang Y, Wang H, Wang Y, Ruderman DL, et al. 2. Scher HI, Fizazi K, Saad F, Taplin ME, Sternberg CN, Miller K, et al. Anterior gradient 2 (AGR2): blood-based biomarker elevated in meta- Increased survival with enzalutamide in prostate cancer after chemo- static prostate cancer associated with the neuroendocrine phenotype. therapy. N Engl J Med 2012;367:1187–97. Prostate 2013;73:306–15.

JANUARY 2017 CANCER DISCOVERY | OF16

RESEARCH ARTICLE Bishop et al.

22. Rapa I, Volante M, Migliore C, Farsetti A, Berruti A, Vittorio Scagliotti 40. Yu X, Cates JM, Morrissey C, You C, Grabowska MM, Zhang J, et al. G, et al. Human ASH-1 promotes neuroendocrine differentiation SOX2 expression in the developing, adult, as well as, diseased pros- in androgen deprivation conditions and interferes with androgen tate. Prostate Cancer Prostatic Dis 2014;17:301–9. responsiveness in prostate cancer cells. Prostate 2013;73:1241–9. 41. Dailey L, Basilico C. Coevolution of HMG domains and homeodo- 23. Arora VK, Schenkein E, Murali R, Subudhi SK, Wongvipat J, Balbas mains and the generation of transcriptional regulation by Sox/POU MD, et al. Glucocorticoid receptor confers resistance to antian- complexes. J Cell Physiol 2001;186:315–28. drogens by bypassing androgen receptor blockade. Cell 2013;155: 42. Lodato MA, Ng CW, Wamstad JA, Cheng AW, Thai KK, Fraenkel 1309–22. E, et al. SOX2 co-occupies distal enhancer elements with distinct 24. Asangani IA, Dommeti VL, Wang X, Malik R, Cieslik M, Yang R, et al. POU factors in ESCs and NPCs to specify cell state. PLoS Genet Therapeutic targeting of BET bromodomain proteins in castration- 2013;9:e1003288. resistant prostate cancer. Nature 2014;510:278–82. 43. Kregel S, Kiriluk KJ, Rosen AM, Cai Y, Reyes EE, Otto KB, et al. Sox2 25. Balbas MD, Evans MJ, Hosfield J,D Wongvipat J, Arora VK, Watson is an androgen receptor-repressed gene that promotes castration- PA, et al. Overcoming mutation-based resistance to antiandrogens resistant prostate cancer. PLoS One 2013;8:e53701. with rational drug design. eLife 2013;2:e00499. 44. Iwafuchi-Doi M, Yoshida Y, Onichtchouk D, Leichsenring M, Driever 26. Azad AA, Volik SV, Wyatt AW, Haegert A, Le Bihan S, Bell RH, et al. W, Takemoto T, et al. The Pou5f1/Pou3f-dependent but SoxB- Androgen receptor gene aberrations in circulating cell-free DNA: independent regulation of conserved enhancer N2 initiates Sox2 biomarkers of therapeutic resistance in castration-resistant prostate expression during epiblast to neural plate stages in vertebrates. Dev cancer. Clin Cancer Res 2015;21:2315–24. Biol 2011;352:354–66. 27. Joseph JD, Lu N, Qian J, Sensintaffar J, Shao G, Brigham D, et al. A 45. Sakaeda M, Sato H, Ishii J, Miyata C, Kamma H, Shishido-Hara Y, clinically relevant androgen receptor mutation confers resistance to et al. Neural lineage-specific homeoprotein BRN2 is directly involved second-generation antiandrogens enzalutamide and ARN-509. Can- in TTF1 expression in small-cell lung cancer. Lab Invest 2013;93: cer Discov 2013;3:1020–9. 408–21. 28. Korpal M, Korn JM, Gao X, Rakiec DP, Ruddy DA, Doshi S, et al. 46. Ishii J, Sato H, Sakaeda M, Shishido-Hara Y, Hiramatsu C, Kamma H, An F876L mutation in androgen receptor confers genetic and phe- et al. POU domain transcription factor BRN2 is crucial for expression notypic resistance to MDV3100 (enzalutamide). Cancer Discov of ASCL1, ND1 and neuroendocrine marker molecules and cell 2013;3:1030–43. growth in small cell lung cancer. Pathol Int 2013;63:158–68. 29. Lee JK, Phillips JW, Smith BA, Park JW, Stoyanova T, McCaffrey EF, 47. Endo T, Yazawa T, Shishido-Hara Y, Fujiwara M, Shimoyamada H, et al. N-Myc drives neuroendocrine prostate cancer initiated from Ishii J, et al. Expression of developing neural transcription factors in human prostate epithelial cells. Cancer Cell 2016;29:536–47. lung carcinoid tumors. Pathol Int 2014;64:365–74. 30. Hubbard GK, Mutton LN, Khalili M, McMullin RP, Hicks JL, Bianchi- 48. Berlin I, Denat L, Steunou AL, Puig I, Champeval D, Colombo S, et al. Frias D, et al. Combined MYC activation and Pten loss are sufficient Phosphorylation of BRN2 modulates its interaction with the Pax3 to create genomic instability and lethal metastatic prostate cancer. promoter to control melanocyte migration and proliferation. Mol Cancer Res 2016;76:283–92. Cell Biol 2012;32:1237–47. 31. Shappell SB, Thomas GV, Roberts RL, Herbert R, Ittmann MM, 49. Boyle GM, Woods SL, Bonazzi VF, Stark MS, Hacker E, Aoude LG, Rubin MA, et al. Prostate pathology of genetically engineered mice: et al. Melanoma cell invasiveness is regulated by miR-211 suppres- definitions and classification. The consensus report from the Bar sion of the BRN2 transcription factor. Pigment Cell Melanoma Res Harbor meeting of the Mouse Models of Human Cancer Consortium 2011;24:525–37. Prostate Pathology Committee. Cancer Res 2004;64:2270–305. 50. Pinner S, Jordan P, Sharrock K, Bazley L, Collinson L, Marais R, et al. 32. Dominguez MH, Ayoub AE, Rakic P. POU-III transcription factors Intravital imaging reveals transient changes in pigment production (Brn1, Brn2, and Oct6) influence neurogenesis, molecular identity, and Brn2 expression during metastatic melanoma dissemination. and migratory destination of upper-layer cells of the cerebral cortex. Cancer Res 2009;69:7969–77. Cereb Cortex 2013;23:2632–43. 51. Bae KM, Su Z, Frye C, McClellan S, Allan RW, Andrejewski JT, et al. 33. Beltran H, Prandi D, Mosquera JM, Benelli M, Puca L, Cyrta J, et al. Expression of pluripotent stem cell reprogramming factors by pros- Divergent clonal evolution of castration-resistant neuroendocrine tate tumor initiating cells. J Urol 2010;183:2045–53. prostate cancer. Nat Med 2016;22:298–305. 52. Jia X, Li X, Xu Y, Zhang S, Mou W, Liu Y, et al. SOX2 promotes 34. Kaplan-Lefko PJ, Chen TM, Ittmann MM, Barrios RJ, Ayala GE, Huss tumorigenesis and increases the anti-apoptotic property of human WJ, et al. Pathobiology of autochthonous prostate cancer in a pre- prostate cancer cell. J Mol Cell Biol 2011;3:230–8. clinical transgenic mouse model. Prostate 2003;55:219–37. 53. Lin F, Lin P, Zhao D, Chen Y, Xiao L, Qin W, et al. Sox2 tar- 35. Tarhini AA, Edington H, Butterfield LH, Lin Y, Shuai Y, Tawbi H, gets cyclinE, p27 and survivin to regulate androgen-independent et al. Immune monitoring of the circulation and the tumor microen- human prostate cancer cell proliferation and apoptosis. Cell Prolif vironment in patients with regionally advanced melanoma receiving 2012;45:207–16. neoadjuvant ipilimumab. PLoS One 2014;9:e87705. 54. Graham V, Khudyakov J, Ellis P, Pevny L. SOX2 functions to maintain 36. Lin D, Wyatt AW, Xue H, Wang Y, Dong X, Haegert A, et al. High neural progenitor identity. Neuron 2003;39:749–65. fidelity patient-derived xenografts for accelerating prostate cancer 55. Catena R, Tiveron C, Ronchi A, Porta S, Ferri A, Tatangelo L, et al. discovery and drug development. Cancer Res 2014;74:1272–83. Conserved POU binding DNA sites in the Sox2 upstream enhancer 37. Grasso CS, Wu YM, Robinson DR, Cao X, Dhanasekaran SM, Khan regulate gene expression in embryonic and neural stem cells. J Biol AP, et al. The mutational landscape of lethal castration-resistant Chem 2004;279:41846–57. prostate cancer. Nature 2012;487:239–43. 56. Ugolkov AV, Eisengart LJ, Luan C, Yang XJ. Expression analysis of 38. Taylor BS, Schultz N, Hieronymus H, Gopalan A, Xiao Y, Carver BS, putative stem cell markers in human benign and malignant prostate. et al. Integrative genomic profiling of human prostate cancer. Cancer Prostate 2011;71:18–25. Cell 2010;18:11–22. 57. Sholl LM, Long KB, Hornick JL. Sox2 expression in pulmonary non- 39. Barton BE, Murphy TF, Shu P, Huang HF, Meyenhofer M, Barton A. small cell and neuroendocrine carcinomas. Appl Immunohistochem Novel single-stranded oligonucleotides that inhibit signal transducer Mol Morphol 2010;18:55–61. and activator of transcription 3 induce apoptosis in vitro and in vivo 58. Morise M, Hishida T, Takahashi A, Yoshida J, Ohe Y, Nagai K, et al. in prostate cancer cell lines. Mol Cancer Ther 2004;3:1183–91. Clinicopathological significance of cancer stem-like cell markers in

OF17 | CANCER DISCOVERY JANUARY 2017 www.aacrjournals.org

BRN2 Is an AR-Suppressed Driver of NEPC RESEARCH ARTICLE

high-grade neuroendocrine carcinoma of the lung. J Cancer Res Clin resistant and MDV3100-resistant prostate cancer growth in vitro and Oncol 2015;141:2121–30. in vivo. Mol Cancer Ther 2013;12:567–76. 59. Laga AC, Lai CY, Zhan Q, Huang SJ, Velazquez EF, Yang Q, et al. 61. Wyatt AW, Mo F, Wang K, McConeghy B, Brahmbhatt S, Jong L, et al. Expression of the embryonic stem cell transcription factor SOX2 in Heterogeneity in the inter-tumor transcriptome of high risk prostate human skin: relevance to melanocyte and Merkel cell biology. Am J cancer. Genome Biol 2014;15:426. Pathol 2010;176:903–13. 62. Goodall J, Wellbrock C, Dexter TJ, Roberts K, Marais R, Goding CR. 60. Kuruma H, Matsumoto H, Shiota M, Bishop J, Lamoureux F, Thomas The Brn-2 transcription factor links activated BRAF to melanoma C, et al. A novel antiandrogen, Compound 30, suppresses castration- proliferation. Mol Cell Biol 2004;24:2923–31.

JANUARY 2017 CANCER DISCOVERY | OF18

The Master Neural Transcription Factor BRN2 Is an Androgen Receptor−Suppressed Driver of Neuroendocrine Differentiation in Prostate Cancer

Jennifer L. Bishop, Daksh Thaper, Sepideh Vahid, et al.

Cancer Discov Published OnlineFirst October 26, 2016.

Updated version Access the most recent version of this article at: doi:10.1158/2159-8290.CD-15-1263

Supplementary Access the most recent supplemental material at: Material http://cancerdiscovery.aacrjournals.org/content/suppl/2016/10/26/2159-8290.CD-15-1263.DC1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerdiscovery.aacrjournals.org/content/early/2016/12/18/2159-8290.CD-15-1263. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.